SANTA CLARA, CALIFORNIA / RankWire.AI / – Nvidia is planning to implement price increases exceeding 15% for numerous AI server configurations slated for shipment in early 2027. These adjustments primarily impact systems based on Vera Rubin and Grace Blackwell technologies. The final increase percentages vary depending on factors such as chip generation, memory capacity, and system design. Nvidia has not announced a unified companywide price hike applicable to all server models. Instead, manufacturers responsible for assembling AI systems have communicated revised prices to their major data center clients.
Microsoft, Google, and Oracle are among the leading cloud service providers purchasing large quantities of accelerated computing hardware. Their data centers deploy AI servers for tasks including model training, inference, and cloud-based services. Throughout 2026, memory has emerged as one of the most significant cost drivers across these systems. Modern AI servers typically combine GPUs with high-bandwidth memory, server DRAM, storage solutions, and rapid networking components. The high demand for these components has kept supply tight in several memory market sectors.
TrendForce forecasts that contract prices for conventional DRAM are expected to rise by 13% to 18% during the third quarter of 2026. Similarly, NAND Flash contract prices are projected to increase by 10% to 15% over the same period. Server DRAM, in particular, remains constrained as memory manufacturers shift more capacity toward AI and data center applications. Elevated memory costs have driven up the expenses involved in building advanced computing systems. These price increases are a key element shaping the pricing landscape for next-generation AI servers.
Memory Pricing Pressures Persist in AI Infrastructure
In 2026, Nvidia reports that Vera Rubin entered full production with system manufacturers and supply-chain partners. Systems utilizing the platform are scheduled to become available in the latter half of the year. Rubin integrates the Vera CPU and Rubin GPU with NVLink 6 and several networking technologies, targeting large-scale AI workloads in cloud and hyperscale data centers. It succeeds Grace Blackwell as Nvidia’s latest rack-scale computing architecture.
Meanwhile, Grace Blackwell remains a fundamental platform within existing AI data center deployments. The GB200 NVL72 system connects 36 Grace CPUs with 72 Blackwell GPUs inside a liquid-cooled rack. Nvidia designed this platform to function as a unified NVLink-based computing domain. Price adjustments associated with these systems vary based on hardware configurations rather than following a fixed percentage. Factors such as memory size, processor generation, and rack design all influence the final server costs.
Demand for Servers Maintains Tight Memory Supply Conditions
As artificial intelligence demand increases, memory manufacturers have shifted more production toward server and high-performance products. TrendForce indicates that this transition has reduced supply availability for some PC and consumer memory segments. Data center operators have continued purchasing substantial volumes of server memory through 2026. The research firm predicts that server DRAM availability will remain limited into 2027, with demand outpacing new supply. This environment continues to influence component costs across AI infrastructure.
Following another quarter of record data center revenue, Nvidia is entering this pricing cycle. The company reported fiscal first-quarter revenue of $81.6 billion for the period ending April 26, 2026. Data Center revenue alone reached $75.2 billion, marking a 92% increase from the same quarter a year earlier. Nvidia has also projected second-quarter revenue of $91 billion, with a margin of plus or minus 2%. The company is set to announce its fiscal second-quarter results on Aug. 26, offering the latest insight into its financial performance.